pubs.acs.

org/cm Published on Web 07/15/2010 r2010 American Chemical Society

4406 Chem. Mater. 2010, 22, 44064413
DOI:10.1021/cm100937e
Hydrophobic, Antireflective, Self-Cleaning, and Antifogging Sol-Gel
Coatings: An Example of Multifunctional Nanostructured Materials for
Photovoltaic Cells
Marco Faustini,

Lionel Nicole,

Cedric Boissiere,

Plinio Innocenzi,

Clement Sanchez,

and David Grosso*

,

Chimie de la Mati ere Condens ee de Paris, UMR UPMC-CNRS 7574, Universit e Pierre et Marie Curie
(Paris 6), Coll ege de France, 11 place Marcelin Berthelot, 75231, Paris, France, and

Laboratorio di Scienza
dei Materiali e Nanotecnologie (LMNT), CR-INSTM, Universit a di Sassari, Palazzo Pou Salid Piazza
Duomo 6, 07041 Alghero SS, Italy
Received April 3, 2010. Revised Manuscript Received June 1, 2010
Antireflective, photocatalytic (self-cleaning), water repellent, and high water-wetting (anti fogging)
properties were combined for the first time into a sol-gel coating deposited onto glass substrates.
Such an original multifunctional coating was obtained by sol-gel liquid deposition of two successive
oxide layers. The first coating is composed of a hybrid methyl-functionalized nanoporous SiO
2
material that exhibits high transparency, high water resistance, close to null water adsorption, and
fairly high mechanical stability (transversal Young Modulus: 1.5 GPa). Thickness and refractive index
can be controlled by selecting proper chemical and processing conditions so as to adjust the anti-
reflectivity properties. The second layer is an ultrathin crystalline TiO
2
nanoperforated layer that was
deposited on top of the previous antireflective layer. Its thickness and refractive index were adjusted
around 12 nm and n 1.8 respectively. This hard TiO
2
top layer acts as a protecting barrier toward
mechanical aggressions and assures high water wetting (antifogging) and photocatalysis (self-
cleaning) at the surface. Indeed, this bilayer systemshows an excellent capability to photodecompose
organic species that were adsorbed into the Anti-Reflective (AR) layer porosity. We show that the
decomposition of model pollutants takes place inside and on the surface of the layer, which is likely
due to diffusion of the pollutants toward TiO
2
and/or diffusion of the radical species toward the
pollutants. These systems are easy to produce on a large scale at lowcost and exhibit high mechanical
and chemical durability. They are thus serious candidates to be used as antireflective, self-cleaning
coatings for photovoltaic cells.
1. Introduction
Many efforts have been dedicated to the development
of novel technologies associated to photovoltaic cells with
the aimof increasingthe energetic yields of suchgreenenergy
production devices. Whatever the type of technology to
photogenerate electricity (silicon based or Dye-sensitive
Solar Cells (DSC)), there is a need to obtain sustainable
optical coatings to allow the maximum solar emission to
reach the exiton generating semiconductor junction. Apply-
ing an Anti-Reflective (AR) layer, with an optimized refrac-
tive index and thickness, on top of a glass top cover of the
device can partly solve this issue.
1-3
Most of the time, the
coatings are primarily porous silica-based material offering
tunable refractive indexandthickness withexcellent adhesion
to the glass surface.
4,5
Unfortunately, pure SiO
2
porous
coatings are not stable in water since they are sufficiently
soluble to be progressively dissolved by rain. This process
is augmented by the presence of the porosity that drama-
tically increases the surface area of the material.
6
In addi-
tion, these pores have hydrophilic surfaces and are small
enough to be filled-up with water at relatively low humi-
dity through capillary condensation leading to a dramatic
effect on the optical properties and thereby increasing the
refractive index and provoking a significant reduction of
the antireflective properties. Meanwhile, the presence of
accessible pores in the nanometer range favoring the
adsorption of pollutants from the out-door environment;
the latter causes higher adsorption of contaminants that
again lead to a rise of the AR layer refractive index and
thus to the loss of its optical function. Water up-take can
be considerably reduced by making the pores hydropho-
bic through the use of alkyl functionalized silica precursors.
7
*To whom correspondence should be addressed. E-mail: david.grosso@
upmc.fr. Fax: 33 (0) 1 44 27 47 69. Phone: 33 (0) 1 44 27 15 30.
(1) Kermadi, S.; Agoudjil, N.; Sali, S.; Tala-Ighil, R.; Boumaour, M.
Thin Films Porous Mater. 2009, 609, 221.
(2) San Vicente, G.; Morales, A.; Gutierrez, M. T. Thin Solid Films
2001, 391, 133.
(3) Wang, X.; Shen, J. J. Sol-Gel Sci. Technol. 2009, online published
(4) Das, S.; Roy, S.; Patra, A.; Kumar Biswas, P. Mater. Lett. 2003, 57,
2320.
(5) Wu, G.; Wang, J.; Shen, J.; Yang, T.; Zhang, Q.; Zhou, B.; Deng,
Z.; Bin, F.; Zhou, D.; Zhang, F. J. Non-Cryst. Solids 2000, 275, 169.
(6) Chowdhury, S. R.; Peters, A. M.; Blank, D. H.A.; ten Elshof, J. E.
J. Membr. Sci. 2006, 276, 276.
Article Chem. Mater., Vol. 22, No. 15, 2010 4407
On the other hand, contamination cannot be avoided
except if the porosity is made non-accessible to pollutants
by applying a hermetic barrier layer. Another alternative
solution would be to integrate a photocatalyst species to
the optical layer whose role would be to clean the porosity
through in situ decomposition of the potential organic
contaminants upon natural sunlight activation. One of
the most efficient materials for such a purpose is TiO
2
anatase which exhibits a total cutoff at wavelength below
350 nm. It has already proven its efficiency in self-clean-
ing window glasses.
8,9
However, problems exist because
of its high refractive index (i.e., n = 2.5), which will
dramatically reduce the AR properties of the layer. To
reduce n, TiO
2
can be nanoparticles embedded into the
AR layer network. In this case, the porosity must be
increased so as to counter balance the increase in material
optical density, inducing a mechanical weakening and a
greater contamination through adsorption in the pores.
The second possibility is to successively deposit both
materials in the desired structure to obtain independent
bilayer systems. Combination of TiO
2
and SiO
2
in the
same coating either under mixed phases
10
or under in-
dependent stack layers
2,3,5,11
obtained through physical
deposition or sol-gel liquid deposition techniques has
recently been reported for different types of optical
applications. None of these solutions combined hydro-
phobicity of the inner porosity together with photocata-
lytic activity located only at the surface, and a perfect
control of both layer porous structure and refractive
index. In the present work, we describe the preparation,
the structure, and the properties of bilayer systems com-
posed of an ultrathin Photo-Active (PA) TiO
2
coating
onto the surface of an Anti-Reflective (AR) hybrid silica-
based layer. The thickness of the top TiO
2
PA layer must
be extremely thin, and its refractive index must be as close
as possible to the refractive index of the AR layer so as to
preserve as much as possible the optical properties. In
addition, it is necessary to reduce the thickness of the AR
layer to compensate for the presence of the PA layer.
The AR layer is constituted of a calcined methyl-
functionalized SiO
2
matrix, exhibiting hydrophobic poro-
sity with pores in the range of several nm in diameter. It
has been prepared by dip-coating glass substrates into a
sol-gel solution containing partly hydrolyzed and con-
densed Tetraethoxysilane (TEOS) and Methyltriethoxy-
silane (MTEOS) precursors in the presence of (PEO-PPO-
PEO) Pluronic F127 block copolymer porogene. After
thermal stabilization, the PAlayer was deposited onto the
previous layer by the same liquid deposition method, but
using TiCl
4
precursor and PB-b-PEO block copolymer.
Thicknesses were adjusted between 50 and 160 nm for the
AR layers and around 15 nm for the PA layer.
12
The re-
fractive index of the AR layer could be adjusted between
1.16 and 1.33 by varying the F127 content, but it was fixed
here at around 1.22 (see Experimental Section). The refrac-
tive index of the PA coating was reduced to around 1.8
through the creation of homogeneous and ordered void
nanoperforations into the material. The scheme of the bi-
layer system is shown in Figure 1. The top layer could be
seenas anultrathinmembrane containing20nmindiameter
perforations organized into a hexagonal structure.
13
Struc-
tural and chemical characterizations were performed using
Environmental Ellipsometric Porosimetry (EEP) techni-
ques, Grazing Incidence X-ray analyses at low and wide
angles (GI-SAXS and GI-WAXS), electronic and near-
field microcopies, FTIR, and UV-visible transmittance.
The bilayer system shows close to 4% increase in trans-
mittance compare to the bare substrate at adjustable wave-
length range, excellent water repellence, good mechanical
properties, self-cleaning properties, and a stability over
9 months immersion into water (longest tested period).
These properties match perfectly with the requirements
for the final cover of out-door solar cells; however, it also
can be applied for other types of optical devices requiring
inner hydrophobicity, anti fogging because of high sur-
face energy, antireflectivity, and photo activity integrated
properties.
2. Experimental Section
Chemicals. Absolute ethanol was purchased from Normapur
while TiCl
4
, TEOS (TetraEthOxySilane) and MTEOS (Methyl-
TriEthOxySilane) precursors were purchased from Aldrich. F127
Pluronic (EO
106
-PO
70
-EO
106
) and P3017-BdEO PB-b-PEO (poly-
butadiene-b-polyethyleneoxide, MWPB=5500g mol
-1
, MWPEO=
5000 g mol
-1
) were purchased from Aldrich and Polymersource
respectively.
Anti Reflective Films. AR films were prepared from solutions
composed of TEOS/MTEOS/F127/HCl/H
2
O/EtOH with respec-
tive molar ratio of 0.5:0.5:0.006:0.005:5:40. MTEOS and TEOS
were first dissolved in the solution composed of EtOH, HCl
(2M) and H
2
O before addition of F127; the solution was stirred
for at least 2 h before deposition. Films were prepared on silicon
and glass microscope slide substrates (n = 1.49 at 700 nm) by
dip coating at room temperature and at low relative humidity
(RH<5%). The thickness of ARfilms was tuned using different
Figure 1. Scheme of the bilayer systemcomposed of the thick ARunder-
neath layer covered by the nanoperforated ultrathin PA layer.
(7) Krzyzak, M.; Frischat, G. H.; Hellmold, P. J. Sol-Gel Sci. Technol.
2007, 41, 99.
(8) Kesmez, O.; Erdem Camurlu, H.; Burunkaya, E.; Arpac, E. Sol.
Energy Mater. Sol. Cell. 2009, 93, 1833.
(9) Zhao, X.; Zhao, Q.; Yu, J.; Liu, B. J. Non-Cryst. Solids 2008, 354,
1424.
(10) Brinker, C. J.; Harrington, M. S. Sol. Energy Mater. J. 1981, 5, 159.
(11) Zhang, X. T.; Fujishima, A.; Jin, M.; Emeline, A. V.; Murakami, T.
J. Phys. Chem.B 2006, 110-50, 25142.
(12) Faustini, M.; Louis, B.; Albouy, P. A.; Kuemmel, M.; Grosso, D.
J. Phys. Chem.C 2010, 114, 7637.
(13) Kuemmel, M.; Allouche, J.; Nicole, L.; Boissiere, C.; Laberty, C.;
Amenitsch, H.; Sanchez, C.; Grosso, D. Chem. Mater. 2007, 19,
3717.
4408 Chem. Mater., Vol. 22, No. 15, 2010 Faustini et al.
withdrawal rates (1, 2, and 4 mm s
-1
) keeping all other dip coat-
ing conditions fixed. After coating, hybrid films were immedi-
atly calcinated underneath a curing IR lamp at 450 C for 10 min.
Photo Active Films. PA were obtained from fresh solutions
composed of TiCl
4
/PB-PEO/H
2
O/EtOH in the following res-
pective molar proportions: 1:0.002:42:160. A first solution A was
prepared by mixing the PB-b-PEO in the water and in 3/4 of the
amount of ethanol. Solution Awas aged at 70 Cfor 2 h until the
complete dissolution of the PB-b-PEO and then cooled down at
room temperature. A second solution B containing TiCl
4
and
the remaining ethanol were added to the cooled solution A before
being stirred at room temperature for 30 min. PA films were
deposited on top of the AR layers by dip coating using a with-
drawal speed of 1.5 mm s
-1
and at a temperature of 40 C, and
relative low humidity. The as-formed bilayer system was then
heated at 500 C for 5 min to ensure crystallization of the TiO
2
top layer. Film samples were labeled as ARX, for the Anti Ref-
lective single layer systems and ARXPA for the Anti Reflective
and Photo Active bilayer systems. X stands for the dip-coating
withdrawal speed in mm s
-1
used to prepare the AR layer. Single-
side coated samples were obtained by cleaning one side of the
substrate prior to heat treatment using a dry tissue imbibed with
ethanol.
Film structure was investigated using Field Emission Gun
Scanning Electron Microscopy (SEM-FEG) Zeiss Ultra 55,
Transmission Electron Microscopy (TEM) JEOL Jem 100CX,
and Atomic Force Microscopy (AFM) Veeco DI-CPII in non-
contact mode using MPP11123 phosphorus doped silicon probes
from Veeco. Grazing Incidence Small-Angle X-ray Scattering
(GISAXS-Rigaku S-max 3000 equipped with a microfocus source
=0.154 nm and a 2D Gabriel type detector place at 1480 mm
fromthe sample) was used to assess the structure of the films and
the period of the array of PA layer nanoperforations at an angle
of incidence of 0.21. The transmitted and specular reflected
beams were masked by a vertical beam-stop. Diffraction pat-
terns were analyzed using Igor software.
14
Ellypsometry mea-
surements were performedona UV-visible (from240 to1000 nm)
variableanglespectroscopicellipsometer (VASE-2000UWoollam),
and the data analyses were performed with the Wvase32 soft-
ware using Cauchy models for both layers. EEPwas investigated
through capillary condensation of water or toluene into the
porosity using an atmospheric control chamber designed by
SOPRA-LAB.
19
Photoactivity of the systems was investigated
by following the refractive index of the layer, contaminated with
Lauric acid, uponUVirradiation. Lauric acidwas chosenas model
pollutant because it contains hydrophilic and hydrophobic
groups found in most organic species. In addition, its carboxy-
late head and its hydrophobic alkyl tail allowstrong interactions
to form independently with the TiO
2
(Lewis acid) surface and
the hydrophobic inner porosity of the underneath layer, respec-
tively. Finally, if one wants to record the mass loss that is only
due to photcatalysis with the present elliposmetry setup, the orga-
nic pollutant must absolutely be poorly volatile. The porosity
was filled up by impregnation with Lauric acid from a concen-
trated solution in ethanol. The Lauric acid was deposited by dip-
coating at 2 mms
-1
the sample into the solution while the filling
of the pores was allowed by impregnation inside the porosity
because of evaporation-induced capillary filling. The extra
Lauric acid deposited on the surface was eliminated using a soft
tissue. In this conditions, ellipsometry analysis revealed that the
whole porosity was filled with the pollutant while no pollutant
layer remains on the surface. The polluted sample was then pla-
ced in air at 5 cm under a UV-lamp (365 nm/24W), and optical
density was recorded every minute using the previous ellipso-
metry conditions over a period of 3 days. The UV/vis absorption
spectra were recorded by using a UVIKON XL SECOMAM
(UVK-Lab) spectrometer directly on the one side coated glass
samples. The background was recorded in air only.
Fourier TransformInfrared (FTIR) analysis has been perfor-
med using a Bruker Vertex70 V spectrophotometer. The optical
bench and the sample compartment have been kept in vacuum
during the measurement at a pressure lower than 0.5 hPa. The
measurements in the middle infrared region have been done
using a Globar source, a KBr beamsplitter, and a RT-DLaTGS
detector averaging 256 scans with 4 cm
-1
of resolution. The mea-
sures in the far-infrared (FIR) region have been performed using
a Globar source, a Si beamsplitter, anda RT-DTGS-FIRdetector.
The spectra have been recorded in transmission, in the 600-
100 cm
-1
range by averaging 32 scans with 4 cm
-1
of resolution.
A silicon wafer has been used as substrate to measure the back-
ground; the baseline has been calculated by a rubberband algo-
rithm (2 iterations, 64 baseline points) (OPUS 7.5 software).
3. Results and Discussion
Figure 2 displays the SEM cross-section image of a typi-
cal AR2PA (see Experimental Section for detailed prepa-
ration) bilayer system where both AR and PA layers can
be independently identified. Indeed, the AR underneath
layer has a thickness of several tens of nanometers (difficult
to precisely assess as a result of the tilting angle) and is com-
posed of randomly ordered pores of around a few nano-
meters indiameter. The topTiO
2
layer has athickness around
10 nm and presents the characteristic nanoperforations
of 20 nm in diameter organized into an hexagonal-type
compact structure. An ellipsometry investigation of the
AR2PAbilayer systemconfirms that both layers could be
independently fitted between 400 and 1000 nm using a non-
absorbing Cauchy layer model. Thickness and refractive
index at 700 nm were found to be respectively 82 nm and
1.22 for the ARlayer and 16 nmand 1.75 for the PAlayer,
confirming the presence of porosity in both layers. Optical
properties of the AR layer were similar before and after
the deposition of the PA layer, suggesting that no Ti-oxo-
cluster entered the AR layer porosity upon the second
Figure 2. SEM image of the AR2PA bilayer system composed of the
thick AR underneath layer covered by the nanoperforated ultrathin PA
layer (scale bar = 100 nm).
(14) Babonneau, D.; Camelio, S.; Lantiat, D.; Simonot, L.; Michel, A.
Phys. Rev.B 2009, 80, 155446.
Article Chem. Mater., Vol. 22, No. 15, 2010 4409
dip-coating process. This impermeability could be attri-
buted to the hydrophobic porosity and to the lowconnec-
tivity with the environment (further confirmed by ellipso-
metry) and also to the fact that PB-b-PEO/Ti-oxo-cluster
hybrid micelles are too large to diffuse inside the porosity.
Indeed, PB-b-PEO in EtOH/H
2
O media forms micelles
with the PB in the core and PEO at the shell. Polar PEO
units are hydrophilic and have the ability to form strong
coordination bonds with Ti (IV) metallic center through
interaction between the free electron pairs of Oatoms and
the vacant d-orbitals of Ti atoms (Lewis acid/base pair).
The Ti-oxo-clusters are then expected to remain in the
PEO environment together with water. Recent studies
15
show that hybrid Ti-oxo-clusters micelles from a similar
PB-b-PEO have an hydrodynamic diameter around 50 nm
which are much larger than the pores of the ARlayer. They,
and their associated coordinated titanium intermediates,
do not penetrate the ARlayer and remainat the surface to
form the nanoperforated PA coating upon thermal treat-
ment. The topography of the bilayer system surface has
been also confirmed by TEM, SEM, and AFM investiga-
tions as shown in Figures 3.
The TEM image (Figure 3a) shows the TiO
2
hexagonal
network, which is darker because of its higher density.
The SiO
2
-based layer can be seen through the perforations
in light gray color. No high resolution or EDX analysis
could be performed because of the diffusion of the elec-
tronic beamthrough the AR layer. The SEM-FEG image
of the surface (Figure 3b) confirms the structure observed
by TEM. The smaller holes, that are visible through the
perforations, belong to the underneath AR layer. The per-
forated topography of the surface is confirmed by AFM
topography (Figure 3c). An additional analysis of the
AR2PAbilayer systemstructure is obtained by GI-SAXS
and GI-WAXS. Figure 4 displays the GI-SAXS 2D
pattern and the GI-WAXS diagram obtained after radial
integration of the exposed image plate. In the GI-SAXS
pattern, thenanoperforationorganizationinto2Dhexagonal-
like structure leads to the presence of the two character-
istics diffraction rods at a d-spacing of 28 nm (together
with their faint harmonic as seen onthe intensity profile in
the y direction (inset Figure 4a)) which corresponds to the
(10) planes. The rods are created by the confinement of
the diffraction waves into the PA thin layer, and the fact
that they are elongated in the vertical direction (normal to
the surface) suggests the presence of a single layer of
nanoperforation motifs distributed with a similar trend
on the whole irradiated surface of the sample. The porous
texture of the bottom AR layer should provide a faint
diffusion ring at greater q-value (marked by white dashed
line on the pattern). Such a low intense signal is, however,
not observed even on the single AR layer as a result of
the poor degree of pore ordering. The WAXS diagram in
Figure 4b presents a wide and low intense diffraction
signal centered at 2=24.8, confirming the crystallinity
of the TiO
2
membrane. Deducing the particle size by app-
lying the Scherrer formula to this diffraction is not appro-
priate because of the grazing incidence geometry by which
the whole width of the sample is irradiated and partici-
pates in the construction of the diffraction. A diffraction
band, corresponding to the more intense (101) peak of the
TiO
2
anatase, is created instead of a real exploitable peak
(see Figure 4 inset). One must point out here that irradi-
ated films that are dimensionally longer lead to a wider
diffraction band. Here a 7 mm wide sample film and a
distance of 60 mm from the detector was used. It seems
that the signal is composed of two overlapping but slightly
shifted peaks. The diffraction band contains two reinfor-
cements below and above 2 =24.8 that correspond to
the same anatase phase but are likely due to the higher
quantity of material present at the edge of the sample
because of typical edge dewetting effects intrinsically
associated to the liquid deposition techniques. The low
intensity is due to the very lowquantity of TiO
2
present on
the surface. Crystallization of the PA layer into anatase
has been confirmed by the far-infrared absorption spec-
tra in the 510-200 cm
-1
range (see Figure 5). The spec-
trum is formed by two intense and well-defined absorp-
tion bands, peaking around 436 and 267 cm
-1
. These
vibrational modes are due to transverse optical (TO) E
u
phonons in tetragonal anatase with two TiO
2
units per
primitive cell.
16,17
Figure 3. (a) TEM, (b) SEM, and (c) AFM images of the AR2PA bilayer system showing the presence of the 20 nm in diameter nanoperforations at the
surface.
(15) Kuemmel, M.; Sm

att, J.; Boissiere, C.; Nicole, L.; Sanchez, C.;

Linden, M.; Grosso, D. J. Mater. Chem. 2009, 19, 3638.
(16) Gonzalez, R. J.; Zallen, R.; Berger, H. Phys. Rev. B 1997, 55, 7014.
(17) Pecharroman, C.; Grac

ia, F.; Holdago, J. P.; Ocana, M.; Gonzalez-

Elipe, A. R.; Bassas, J.; Santiso, J.; Figueras, A. J. Appl. Phys. 2003,
93, 4634.
4410 Chem. Mater., Vol. 22, No. 15, 2010 Faustini et al.
As expected and already observed with the present
nanoperforated TiO
2
top layer in a previous study,
18
the
deposition of a drop of water on top of the bilayer surface
leads to the complete wetting with a contact angle inferior
to 10, confirming the antifogging property of the layer.
Such a property is due to the high surface energy of TiO
2
.
EEPis a recent and very efficient technique to assess the
porosity of an optical coating by probing the adsorption/
desorption and capillary condensation of selected vapors
inside the film. First, the experiment was performed on
the AR layer using water vapor. The sample was exposed
to increasing and decreasing relative vapor pressure (0<
P/P
0
< 1), and the evolution of refractive index was re-
corded every P/P
0
=0.02. The refractive index was found
to vary from 1.23 to 1.25 when exposed from 0 to 100%
relative humidity (data not shown), suggesting that less
than 5% of the porosity can be filled up with condensed
water while the rest of the porosity remains empty even at
saturated humidity. As already mentioned previously,
this behavior is likely due to the presence of residual
methyl groups making the surface of the pores hydro-
phobic. The same film was then aged for 9 months in
deionized water, and the same water repellent behavior
was observed. The AR layer refractive index and thick-
ness were not modified by the latter treatment, suggesting
that the methyl-containing silica framework does not
dissolve even after a long period into water. In a second
EEP experiment, non-polar toluene vapor was used as
adsorbate. In this case, the adsorption/desorption iso-
therm shows a toluene capillary condensation into the
pores suggesting that the porosity is open to the atmos-
phere. Figure 6a shows a progressive increase of the refrac-
tive index of the AR layer with increasing toluene vapor
pressure. Capillary condensation for a given P/P
0
value
can be related to a specific pore dimension and morpho-
logy.
19
In this case, the pore size distribution in Figure 6b)
has been obtained from the Kelvin equation using the
model describing interconnected pores (cylindrical model)
and using the physical and chemical characteristics of
toluene.
19
It confirms that a population of pores having
diameters below 10 nm are present in the structure. Such
a finding is in agreement with the microscopic analysis
where pores of various dimensions but below10 nminsize
could be depicted. The desorption branch shows a sudden
decrease of the refractive index below 50% that corres-
ponds to the desorption of the toluene. This is character-
istic of the bottle neck effect, where smaller intercon-
necting windows assures accessibility to the pores. An
important point to underline here is that only non-polar
liquids, such as toluene, can condense inside the pores in
the studied conditions, which proves that the interface of
the pores must be hydrophobic. This hydrophobicity must
be associated with uncompleted thermal decomposition
of the methyl groups on the surface of the pores, which
will be confirmed later by FTIR investigation. For com-
parison, the same filmbut free of methyl groups, prepared
with 100% TEOS as precursor, shows high hydrophili-
city.
19
Mechanical properties of the hybrid AR film were
investigated by analyzing the evolution of film thickness
in the range of toluene relative pressure for which capil-
lary condensation has occurred (P/P
0
>0.6). Using the
Kelvin-Laplace equation given in ref 19 to fit the capillary
stress relaxation of the film, we deduced the transverse
Figure 4. (a) GI-SAXS pattern and (b) GI-WAXS diagram obtained for a typical AR2PA bilayer system.
Figure 5. Far infrared absorption spectrum of a titania monolayer obta-
ined on a AR2PA system.
(18) Jarn, M.; Brieler, F.; Kuemmel, M.; Grosso, D.; Linden, M. Chem.
Mater. 2008, 20, 1476.
(19) Boissiere, C.; Grosso, D.; Lepoutre, S.; Nicole, L.; Brunet-Bruneau,
A.; Sanchez, C. Langmuir 2005, 21, 12362.
Article Chem. Mater., Vol. 22, No. 15, 2010 4411
Young Modulus (E) to be around 1.5((0.3) GPa. As the
maximum film contraction observed upon capillary con-
densation is about 3%of the filminitial thickness, we can
conclude that the mesoporous hybrid AR layer is behav-
ing mechanically very much like mesoporous purely inor-
ganic ceramics.
19
Thus, organic functions introduced into
the silica network are mainly located at the interface of
the silica walls and do not weaken the silica framework
beyond an acceptable point for the AR coating applica-
tion. In addition, the bilayer system was resistant to
cleaning with a soft tissue soaked either by water, ethanol,
or acetone. The complete removal of the bilayer coating
was achieved only after 5 cycles of aging 5 min in a con-
centrated KOH solution followed by rubbing with a dry
tissue, confirming its relatively good mechanical stability.
Figure 7 shows the infrared absorption spectrum in the
4000-900 cm
-1
range of a MTEOS-TEOS hybrid film
that has been treated at 450 C for 10 min and immedi-
ately increased to 500 C for 5 min. The spectrum shows
an intense absorption band peaking at 1050 cm
-1
that is
assigned to Si-O-Si antisymmetric stretching, which
indicates the formation of the silica network. On the other
hand this network appears well condensed because no
signals due to silanols at 910 cm
-1
(Si-OH stretching),
and 3300 cm
-1
(O-H stretching) can be observed. Two
small bands at 3657 and 3765 cm
-1
are assigned to the pre-
sence of tween (geminal) and isolated silanols, respectively,
which are the residual of the condensation of silanol chains.
This is in accordance with previous findings and is a good
indication of the state of the surface upon condensation.
On the other hand, the spectrum gives also some impor-
tant information: no signals due to absorbed water are
present, neither at 1640 cm
-1
nor at 3200 cm
-1
(O-Hstret-
ching), and a strong sharp signal at 1280 cm
-1
indicates
the presence of CH
3
groups in the film. The FTIR data
indicates therefore that the film structure is formed by a
well condensed silica network which is modified via the
covalent Si-C bonds with methyl groups mainly located
at the surface of the pores.
20-22
The fact that methyl
groups were not fully decomposed at such high tempe-
rature is not yet fully understood but is likely due to the
confinement and fast thermal treatment. The presence of
these hydrophobic species avoids the absorption of water
or other hydrophilic moieties from the external environ-
ment within the film.
Figure 7. FTIR absorption spectra of a MTES-TEOS hybrid film after
thermal treatment.
Figure 6. (a) Evolution of refractive index of the ARlayer (AR2 sample) upon exposure to increasing and decreasing toluene vapor pressure. (b) Pore size
distribution deduced from the adsorption curve.
Figure 8. Evolution of refractive index with UVirradiation time for both
single AR (AR2 sample) layer and AR-PA bilayer (AR2PA sample)
systems contaminated with Lauric acid.
(20) Innocenzi, P. J. Non-Cryst. Solids 2003, 316, 309.
(21) Malfatti, L.; Kidchob, T.; Falcaro, P.; Costacurta, S.; Piccinini,
M.; Cestelli Guidi, M.; Marcelli, C.; Corrias, A.; Casula, M.;
Amenitsch, H.; Innocenzi, P. Microporous Mesoporous Mater.
2007, 103, 113.
(22) Falcaro, P.; Grosso, D.; Amenistch, H.; Innocenzi, P. J. Phys.
Chem. B 2004, 108, 10942.
4412 Chem. Mater., Vol. 22, No. 15, 2010 Faustini et al.
Photocatalytic efficiency of both single AR layer and
AR2PA bilayer systems was investigated by ellipsometry
in situ during UV irradiation. The hydrophobicity of the
pore permits to fill them up with Lauric acid pollutants.
The contaminated system was then placed below a UV-
lamp, and the variation of refractive index was recorded
every minute by ellipsometry over a period of 3 days (see
Figure 8). The initial thickness and refractive index of the
AR layer before impregnation were 130 nm and 1.23, res-
pectively. After impregnation with Lauric acid, the thick-
ness does not change while the refractive index increases
up to 1.43, suggesting that a fair proportion of the poro-
sity volume was contaminated. While the AR2PA bilayer
system regained its initial porous state after 2 days of
irradiation, the TiO
2
free system has eliminated only less
than 50% of its contamination by natural non-catalyzed
UV decomposition for the same period, confirming the
photocatalytic role of the TiO
2
PA top layer. Even if
the TiO
2
is locatedoutside of the ARlayer, it provokes the
photodecomposition of the Lauric acid molecules located
inside the AR layer by ellipsometry, suggesting that con-
taminant species do not have to be in direct contact with
TiO
2
to be eliminated. In this conditions, the decreasing
of the refractive index is related to the decomposition of
the Lauric acid presumably associated to the formation of
OH
3
and O
2
3
radicals. The latter are produced as a result
of the electronic transfer operating between the TiO
2
semi-
conductor surface and the adsorbed water upon irradia-
tion. The products of the reaction between the radicals
and the Lauric acid become more and more volatile and
eventually evaporate fromthe film. For the bilayer geometry,
one may state that Lauric acid diffuses toward the TiO
2
rich zone (surface), where irradiation is optimal and where
the concentration of free radical is high. However, one
cannot neglect the possibility that even if highly unstable,
free radicals may have a sufficient lifetime to migrate inside
the AR layer to decompose the pollutants.
Antireflective properties of both single AR layer and
both AR-PA bilayer systems were assessed by UV-visible
spectroscopy for different AR layer thickness. The per-
centages of transmitted light (transmittance) of AR, AR-
PA, and substrate systems are plotted in Figure 9 in the
visible wavelength range. In the inset table in Figure 9,
one observes that the refractive index of the AR layer of
the single and bilayer systems are close to 1.23 whatever
the thickness. The PA layer has a refractive index and
thickness close to 1.8 and 15 nm, respectively. As expec-
ted, the single ARcoatings with a refractive index close to
1.23 provide the characteristic increase in transmission of
around 4.5%, with a maximum transmission, correspon-
ding to the first harmonic of the interference, shifting to
higher wavelengths for higher thicknesses.
For the bilayer systems, the transmitted intensity is
considerably decreased at lower wavelengths as a result of
the anatase TiO
2
absorption below 350 nm. On the other
hand, one notices that the antireflective properties are
maintained in the visible range, reaching 3.5% gain at
maximal transmission, even if they are slightly lower than
that of the single AR layer systems. One can also observe
that the maxima are shifted to higher wavelength because
of the increase in optical path resulting from the presence
of the TiO
2
layer. The latter results confirm that the bilayer
Figure 9. UV-visible analysis intransmissionat 90 incident angle of the single layer (plainlines) andbilayer (dottedlines) systems depositedonone side of
the glass microscope slide substrate. Thickness and refractive index (measured by ellipsometry) of each system are reported in the joint table.
Article Chem. Mater., Vol. 22, No. 15, 2010 4413
systems has efficient AR properties, resulting from the
thinness and reduced refractive index of the PA top layer,
and that the wavelength of maximumtransmission can be
adjusted by the thickness of the AR layer.
4. Conclusions
We have confirmedthat antireflectivity, water-repellence,
high wetting behavior (antifogging), photocatalytic acti-
vity (improved self-cleaning ability), and relatively high
mechanical properties can be integrated into the same
sol-gel coating made of a methyl modified SiO
2
and TiO
2
bilayer. The localization of the hard semiconductive TiO
2
ultrathin layer at the surface of the ARhydrophobic silica
layer, provides (i) the surface high energy necessary for
high wetting, (ii) the direct exposure to radiation and
pollutants, and (iii) an improved resistance to abrasion.
Chemical precursors used to produce such materials are
relatively low cost while the dip-coating liquid deposition
process is highly suitable to large-scale production. The
systemshowed no alteration of the structure and the opti-
cal properties after 9 months immersion in water, making
these coatings serious candidates for use as antireflective,
self-cleaning top coating for photovoltaic cells.
Acknowledgment. D. Jalabert (Centre de Microscopie
Electronique dOrleans) is thanked for TEM analyses and
Dr. Vasana Maneeratana for the language corrections. The
authors also acknowledge funding provided by The TER-
AMAGSTOR program, the CNRS and the UPMC.